The proposed research builds on previous and current studies of computer- based remediation for reading-disabled children, with a new comparison of two theoretically and practically different ways of training phoneme awareness. The story reading portion of the training will be similar to previous grants. In each training year of the proposed study, 108 second- to fifth-grade students with reading disabilities will read interesting stories on a computer for 30 minutes daily during their remedial-reading or language-arts class time. They will target difficult words in the stories with a mouse for immediate orthographic and speech feedback. The computer will emphasize important relations between sub-word letter patterns and speech sounds, by segmenting the word into helpful units for the student to sound out, and then by pronouncing each segment in order while highlighting the segment. Previous studies revealed significant gains in phonological decoding, word recognition, and attitude about reading for reading-disabled students trained on the computer compared to similar students receiving normal classroom or remedial reading instruction. However, trained students with relatively lower phoneme awareness had much smaller average gains than those with higher phoneme awareness. Current studies show that supplementary, articulatory-based training in phoneme and phonological awareness on the computer can reduce or erase these deficits and support substantially greater gains in word decoding. Now we propose to compare two different phonological-awareness programs and an untrained control group. Both training conditions will include intensive practice in manipulating phonemes and graphemes in reading and spelling exercises, but one will pretrain articulatory awareness while the other pretrains sound-letter based phoneme awareness, to isolate unique effects of concrete speech-motor awareness. We hypothesize that training in articulatory awareness will be most critical for children with the weakest initial phoneme awareness, while other children with less severe reading disabilities may benefit as much from non-articulatory based training in phoneme awareness and phonological decoding. We will also explore the addition of visual speech cues from an animated mouth and face that coordinates with the synthetic speech. Longitudinal assessment through individual growth-curve analyses will describe subjects' development of reading and language skills during and up to 1 year after training.